AI & bot directory · AI training
GPTBot: what it is, how to verify it and how to block it
GPTBot is run by OpenAI. Crawls content that may be used to train OpenAI's generative AI models.
Key facts
| Operator | OpenAI |
|---|---|
| Category | AI training |
| robots.txt name | GPTBot |
| Follows robots.txt | OK: Yes |
| Can it be verified? | OK: Yes, by IP ranges |
| Official documentation | |
| Registry checked | Nov 11, 2026 |
User agent
Its name in robots.txt is GPTBot.
OpenAI documents this user-agent string:
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.4; +https://openai.com/gptbot
Can you verify GPTBot?
Yes OpenAI publishes the IP addresses GPTBot uses. We download that list every 6 hours: our copy is from Nov 15, 2026 and has 18 entries; OpenAI dated it Sep 22, 2026.
Does GPTBot follow robots.txt?
Yes OpenAI says GPTBot follows robots.txt.
“Disallowing GPTBot indicates a site's content should not be used in training generative AI foundation models.”
How to block or allow GPTBot
To ask it to stay away from the whole site, add this to robots.txt:
User-agent: GPTBot
Disallow: /
To keep it out of one folder only:
User-agent: GPTBot
Disallow: /private/
To allow it while asking every other bot to stay away:
User-agent: GPTBot
Allow: /
User-agent: *
Disallow: /
robots.txt is a request, not a lock. To enforce a block, add a firewall or CDN rule. Because OpenAI publishes GPTBot's addresses, a rule can match them exactly, so a fake GPTBot can't slip through as the real one.
To build one robots.txt for several bots at once, use the robots.txt generator.
How often does GPTBot visit?
OpenAI doesn't publish how often GPTBot visits a site, and iplover doesn't see your traffic. Your own access log is the only reliable count: paste it into the log checker, which counts each crawler's visits and flags fakes, in your browser.
Official documentation
Registry checked Nov 11, 2026. The address lists are downloaded every 6 hours.